Advancing the measurement and use of daily life heart rate using wearable technology: applications to Parkinson's disease

Abstract

Measures of heart rate (HR), including those conducted during resting conditions and in response to physical activity, have been shown to relate to risk of adverse health outcomes and can provide insight into the function of the autonomic nervous system (ANS). However, HR measures have conventionally been conducted at a single point in time and in controlled settings, such as during clinical or laboratory visits. Alternatively, when HR has been measured in uncontrolled, free-living settings, there has been a lack of standardization in the methods used and most have not provided sufficient behavioral context to support the interpretation of HR in daily life. Use of wearable technology, and specifically, a multi-sensor and -device approach, presents a unique opportunity to measure HR continuously in the context of daily life behaviors, and advance understanding of ANS function and context-dependent HR responses. This may be particularly relevant to individuals living with Parkinson’s disease (PD), who can experience ANS dysfunction that may manifest with altered diurnal HR profiles and who can have diverse movement characteristics that may affect the HR response to physical activity.

The overarching goals of this thesis work were to 1) advance the development of standard approaches to assess HR from continuously collected data over multiple days and behavioral contexts, and 2) apply these approaches to advance understanding of ANS function and characterization of daily life physical activity profiles among individuals with PD. All studies were based on a collection protocol that instrumented participants with a portable, chest-worn electrocardiography device and two limb-worn inertial measurement units (wrist and ankle) that continuously collected data across a week-long period in daily life. In the first study, HR was measured across accumulated periods of sedentary behavior and sleep within a cohort of adults without PD (“control”), and the high levels of within-participant variability in HR that were observed within a given resting condition led to recommendations being put forth for standardizing the measurement of HR outcomes during resting conditions in daily life. In the second study, standards suggested in Study 1 were applied and extended upon in a cohort of adults with and without PD, and revealed no differences between cohorts in resting condition-specific HR outcomes (e.g., minimum HR during sleep) or in measures of nocturnal HR dipping in daily life. However, a significant association was observed between nocturnal HR dipping and a clinical measure of ANS dysfunction in PD. In the third study, resting HR was used as a reference when quantifying the relative intensity of the HR response to daily life walking across the same cohort of adults with and without PD. While no between-cohort differences were observed in the average relative intensity of daily life walking, there was considerable variability in the between- and within-participant HR response to continuous, longer walking bouts (≥ 30 seconds), which were not consistently associated with walking cadence or bout duration.

This dissertation provides recommendations for, and illustrates the importance of, adopting standardized approaches to measure resting and activity-dependent HR in daily life. Findings from the studies in this thesis suggest that context-based ambulatory HR monitoring has the potential to improve characterization of daily life behaviors and provide insight into clinically relevant outcomes in individuals with and without PD. Continued research in larger, more heterogenous populations, which includes robust clinical profiling and continues to attend to the importance of high data quality, is necessary to determine generalizability and advance understanding of within- and between-participant drivers of variability in context-dependent HR measures. This type of future work is required before context-dependent HR monitoring can be integrated and evaluated as a tool to support self-management or clinical health care decision making.